Microstructure and texture evolution during cold rolling and annealing of a high Mn TWIP steel

高Mn合金元素TWIP钢微观组织和织构演变

Available online at http://wendang.chazidian.com

ActaMaterialia57(2009)

1512–1524

http://wendang.chazidian.com/locate/actamat

Microstructureandtextureevolutionduringcoldrolling

andannealingofahighMnTWIPsteel

L.Brackea,1,K.Verbekena,b,*,L.Kestensa,c,J.Penninga

a

DepartmentofMaterialsScienceandEngineering,FacultyofEngineering,GhentUniversity(UGent),Technologiepark903,B-9052Ghent,Belgium

b

MicrostructurePhysics,Max-Planck-Institutfu¨rEisenforschung,Max-Planck-Strasse1,40237Du¨sseldorf,Germanyc

MaterialsScienceandEngineeringDepartment,DelftUniversityofTechnology,Mekelweg2,2628CDDelft,TheNetherlands

Received28March2008;receivedinrevisedform25November2008;accepted27November2008

Availableonline7January2009

Abstract

Themicrostructureduringcoldrollingandannealingofalow-stackingfaultenergyausteniticFe–Mn–Calloywasstudiedbymeansofelectronmicroscopy.Thecontributionofbothslipandmicrotwinningtothedevelopmentofabrass-typecold-rollingtexturewasillustrated.Duringsubsequentrecrystallizationannealing,aretainedrollingtexturewasobserved.Itwasshownthatthemechanismbehindthisphenomenonwasbasedonthenucleationandgrowthoftherecrystallizedstatewithoutpreferredorientationinanenerget-icallyrelativelyhomogeneousmicrostructure.

Ó2008ActaMaterialiaInc.PublishedbyElsevierLtd.Allrightsreserved.

Keywords:Austeniticsteels;Deformationtexture;Recrystallizationtexture;Transmissionelectronmicroscopy;Electronbackscatterdi raction

1.Introduction

Thecurrenttrendintheautomotiveindustryistoreducevehicleweightinordertodecreasefuelconsumptionandtheassociatedemissionsofgreenhousegases.Hence,astringentneedhasemergedforhigh-strengthmaterialswithexcellentformability,whichhasledtothedevelopmentofnewalloysingeneral,andsteelsinparticular.Typicalexamplesofnewhigh-strengthsteelgradesarethefer-rite–martensitedualphase(DP)steelsandthemultiphasesteelsexhibitingatransformation-inducedplasticity(TRIP)e ect.DPandTRIPsteelscanreadilyachieveulti-matetensilestrength(UTS)levelsintherange600–1000MPaincombinationwithelongationsintherange

Correspondingauthor.Address:DepartmentofMaterialsScienceandEngineering,FacultyofEngineering,GhentUniversity(UGent),Tech-nologiepark903,B-9052Ghent,Belgium.Tel.:+3292645784;fax:+3292645833.

E-mailaddress:kim.verbeken@ugent.be(K.Verbeken).1

Presentaddress:CorusResearch,DevelopmentandTechnology,P.O.Box10000,1970CAIjmuiden,TheNetherlands.

*

15–25%.Suchpropertiesmakethesesteelsveryattractiveforautomotiveapplicationswherehighstrengthisrequired.Forapplicationsrequiringimprovedformability,e.g.,fortheproductionofcomplexpartsinoneformingoperation,austeniticsteelgradesareaverypromisinggroupofmaterials.Themechanicalbehaviourand,inpar-ticular,thestrainhardening,ofthesesteelslargelydependonthestackingfaultenergy(SFE),becausethismaterialparameterdetermineswhichdeformationmechanismwillbeactivated[1–4].WithalowerSFE,thedeformationmechanismchangesfromslipofperfectdislocationstoslipofpartials,tomechanicaltwinningandeventuallytotrans-formationintohcpe-martensiteand/orbcc/bcta0-martens-ite.Itisimportant,though,tobearinmindthatthegreaterpartoftheplasticslipisaccommodatedthroughslipofdis-locations,regardlessoftheactivationofotherdeformationmechanisms.

ThecriticalvalueoftheSFEforthetransitionfrome-mar-tensiteformationtomechanicaltwinningisofspeci cinterest,astwinningisexpectedtoincreaseboththeUTSandtheuni-formelongation,aphenomenonknownasthetwinning-inducedplasticity(TWIP)e ect[5,6].AccordingtoOhetal.

1359-6454/$34.00Ó2008ActaMaterialiaInc.PublishedbyElsevierLtd.Allrightsreserved.doi:10.1016/j.actamat.2008.11.036

高Mn合金元素TWIP钢微观组织和织构演变

L.Brackeetal./ActaMaterialia57(2009)1512–15241513

[1],twinningoccurswhenanalloyhasaSFE>18mJmÀ2,whiletheformationofe-martensiterequireslowervalues.Allainetal.[2]calculatedthattwinningcanoccurforaSFEbetween12and35mJmÀ2,wherease-martensitecanbe

formedforSFEvalues<18mJmÀ2.TheresultsofRe

´myandPineau[3]indicatethattheminimumSFErequiredformechanicaltwinningis9mJmÀ2,whilee-martensitecanbeformedwhentheSFEdoesnotexceed12mJmÀ2.Takingintoaccountthedi cultiesinreliablydeterminingtheSFEandthedi erentcompositionsofthealloysinthesestudies,theagree-mentbetweenthedi erentliteraturedataisreasonablygood.TheSFEisknowntohaveastrongin uenceonthedevelopmentofthecrystallographictextureduringcoldrolling.High-SFEmaterialssuchasAl,CuandNiatroomtemperature,whichdonotdeformbytwinning,developaCuorpuremetaltexture,whichischaracterizedbyastrongb- bre,withanincreasingintensityfromthebrasscomponent,overtheStothecoppercomponent[7–10].Low-SFEmaterialsthatcantypicallydeformbytwinningdevelopabrassoralloytexturewithalowerintensityonthecoppercomponentandanincreasedoneonthebrassandGosscomponents,i.e.,thefcca- bre.Thedevelop-mentofsuchatextureisoftenattributedtomechanicaltwinning[11,12].Ithasalsobeenreportedthatmechanicaltwinninginitselfisnotresponsibleforthedevelopmentofabrasstexture,butthatitinitiatestheformationofmacro-scopicshearbandsresultinginabrasstexture[13].Abrass-typetextureevolutionhasalsobeenobservedinlow-SFEmaterialsbeforetwinningwasobserved[9].Inthiscase,thetexturedevelopmentwasattributedtomicroscopicshearbands[9].Inlow-SFEmaterials,thepartialdisloca-tionsformingthestackingfault(SF)arewidelyseparated.Thisrestrictsthepossibilitiesforcross-slip,astherequiredperfectscreworientationforacross-slippingdislocationcanonlybeachievedbyrecombinationofthetwopartials.Hence,theslipisrestrictedtotheplaneinwhichtheSForiginated,i.e.,planarslip.Recently,thein uenceoftheplanarityoftheslipsystemsandthelocalizationofslipinfccmetalswithalowSFEhasbeenmodelledbyMirigliaetal.[14].Theyfoundthattheirmodelcouldonlypartiallyaccountfortheexperimentalobservations.Texturestudiesoflow-SFEausteniticsteelsrevealtheimportanceofstrain-inducedtransformationofaustenitetoa0-martensite[15].Thegeneraltendencyoftheaustenitetexturedevelop-mentinthatcaseisstilltheformationofabrass-typetex-ture,evenwhennotwinninghasoccurred.However,theappearanceofstrain-induceda0-martensiteisexpectedtoslowdownthedevelopmentofthebrass-typetexture[16].Whilethetextureevolutionduringcoldrollingoflow-SFEmaterialsisfairlywellunderstood,muchlessclarityexistsontherecrystallizationbehaviour.High-SFEmateri-als,suchasAl,CuandNi,typicallyshowarecrystalliza-tiontexturedominatedbythecubecomponent.Severalstudieslinkthistoaselectivegrowthmechanism[17–21],whileotherspointtoorientednucleation,basedonthelow-storedenergyofthecubeorientation[22,23].Forlow-SFEmaterials,however,therecrystallizationtextures

areevenlessunderstood[8].AstudybyLu¨cke[24]onCu-basedalloyswithdi erentSFEshowedthat,withdecreasingSFE,the{236}h385i(Eulerangles:u1=79°,U=31°,u2=34°)componentbecamethedominanttex-turecomponent.Veryfewdataontherecrystallizationtex-tureoflow-SFEausteniticsteelsareavailableintheliterature.OneexampleistherecentworkbyVercammen[25]onaFe–30Mn–3Si–3AlausteniticsteelwithaSFEof40mJmÀ2inwhichasmallcubecomponentwasobservedafterrecrystallizationannealing.

Theaimofthepresentworkistostudythedeformationbehaviourandtherecrystallizationmechanismofalow-SFEFe–Mn–C-basedTWIPalloy.Therefore,adetailedanalysisofthemicrostructureandtextureevolutionwascarriedout.2.Experimental

Thematerialstudiedforthestudyofthecold-rolledmicrostructureandtextureisafullyausteniticFe–22%MnalloywithconsiderableamountsofCandNtostabi-lizetheausteniteatroomtemperature.TheSFEofthismaterialwas15mJmÀ2[26].Thematerialwascastasan80-kgingotinalaboratoryvacuummeltingandcastingunitunderaprotectiveN2atmosphereof1013hPa.Afterreheatingat1473K,thematerialwashotrolledonalabo-ratorymillfrom20to3mminfourpasses,maintainingthe nishingtemperature>1223Ktoallowforstaticrecrystal-lizationbeforewaterquenching.Samplesofthehot-rolledstripwerecoldrolledwithtotalreductionsrangingfrom5%to50%(cf.Table1).

Therecrystallizationannealingtreatmentswereper-formedinsaltbathfurnacesatdi erenttemperaturesbetween673and998Kfor120swithintervalsof25Kandataconstanttemperatureof998Kforannealingtimesbetween10and120s.Thesamplesforthesetreatmentswereall50%coldrolled.

Therecrystallizedfraction(%ReX)wasderivedfromtheVickershardness(HV3)measurements,whichwereper-formedwithaloadof29.4N(HV3)usingaZwick3212hardnesstester.Atleastthreetestswereperformedperdatapoint.Formeasurementswithaspread>10HV3,atleast vemeasurementsweredoneperdatapoint.TexturemeasurementsbymeansofX-raydi ractionwereper-formedonaSiemensD5000di ractometer,usingMoKaradiation(k=0.070926nm)at50kVand50mA,bymea-suringpole guresandcalculatingorientationdistributionfunctions(ODF).The{100},{110},{211}and{310}incompletepole guresweremeasuredbytheSchultzre ectionmethodusingaEulercradlegoniometer.The

Table1

Overviewoftheappliedrollingstrains:engineeringstrainvalues(e)andlogarithmictruestrainvalues(e).e

0.050.100.200.300.400.50 ¼Intinit

0.05

0.11

0.22

0.36

0.51

0.69

高Mn合金元素TWIP钢微观组织和织构演变

1514L.Brackeetal./ActaMaterialia57(2009)1512–1524

ODFwerecalculatedwiththeMTM-FHMsoftwaredevel-opedbyVanHoutte[27].

AZeissDSM962scanningelectronmicroscopeoper-atedat15kVwasusedformicrostructuralinvestigation.Electronbackscatterdi raction(EBSD)measurementswerecarriedoutontheplaneparalleltotherollingandnormaldirection.Scanswithstepsizes0.1–0.5lmwerecarriedoutonaFEIXL30ESEMequippedwithaLaB6- lament,usingaTSL–OIMEBSDattachment.ThesamplepreparationforXRDandOIMobservationscon-sistedinmechanicalpolishingwith1lmdiamondpaste,followedbymechanicalpolishingwithStruersOPSsolu-tion.Noetchingwasapplied.

TheKernellaveragemisorientation(KAM)wasusedtoillustratetheenergetichomogeneityofthecold-rolledstructure.TheKAMisameasureoflocalmisorientationandisdeterminedasfollows.ForeachpointmeasuredintheEBSDscan,theaveragemisorientationofthatpointanditsneighboursiscalculatedusingTSL–OIMsoftware,undertheconditionthatmisorientationsexceedingsometolerancevalue,e.g.,5°,arediscarded.Thisisnecessarytoexcludethee ectoflargemisorientations,e.g.,duetothepresenceofagrainboundary.

Thepartitioningbetweenrecrystallizedandunrecrystal-lizedregionswasdone,basedonthegrainorientationspread(GOS)parameter,asdeterminedintheEBSDmea-surements.Thisparameterhasalreadybeenusedelsewhereforthesamepurposesinaluminiumalloys[28,29].Thecal-culationoftheGOSisdoneintwosteps.Theaveragecrys-tallographicorientationofallthepointswithinagrainiscalculated rst.TheGOSistheaveragedeviationbetweentheorientationofeachindividualpixelandtheaverageori-entationofthegrain.Whenthisisdoneforeverygrain,aGOSdistributioncanbedeterminedforthecompletemeasurement.

Asdeformedgrainspossessinternalorientationgradi-ents,theyproducehigherGOSvalues.Consequently,thisparameterallowsonetodistinguishbetweendeformedandrecrystallizedgrains.Otherstudies(e.g.,Ref.[30])haveusedtothispurposetheimagequality(IQ)parameter,whichquanti esthesharpnessoftheEBSDpatterns.Itwaspointedout,however,thatclearlydeformedgrains

canstillexhibitareasonablyhighIQfactor,whichmakesthisparameterunsuitablefordeterminingtherecrystallizedfraction.GrainswithanaverageCon denceIndex(CI)<0.1wereexcluded,althoughthishadalmostnoe ectonthefractionsobtained.

Thinfoiltransmissionelectronmicroscopy(TEM)wasperformedonaPhilipsEM420operatedat120kVandonaPhilipsCM200FEGoperatedat200kV.The nalsamplepreparationconsistedindoublejetelectrolyticpol-ishingwithaStruersTenupol5,operatedatavoltageof32V.Theusedelectrolytewasasolutionof93%CH3COOHand7%HClO4atatemperatureof285K.3.Results

3.1.Microstructureinthecold-rolledstate

Fig.1showsmicrographsofthealloybefore(Fig.1a)andaftercoldrolling(e=0.22)(Fig.1b)ande=0.51(Fig.1c).

Thelocalizeddeformationlinesinthecold-rolledspeci-menmaypresumablybeattributedtoextensivetwinningofthealloyduringdeformation,butitwillbeshownlaterthattheselinescanalsobesliplineswithoutanytraceoftwin-ning.Athigherstrains(e=0.51),thelamellardeformationstructurerevealsmacroscopicshearbands,cf.Fig.1c.Thistypeofshearbandsisadirectconsequenceofexcessivemechanicaltwinning[13,31].Itwasnotpossibletoverifywhethermorepronouncedmacroscopicshearbandingoccursathigherrollingstrains,becausethemaximumforceoftheusedlaboratoryrollingmillwasnotsu cienttoreducethethicknessofthesamplefurther.

Fig.2showsEBSDresultsforthreesimilarlookingdeformedareasaftercoldrollingtoastrainofe=0.51(1–3)andofanotherareacoldrolledtoe=0.22(4).

Theseareaswereselectedfromafullmeasurementbymeansofthepost-processingTSL–OIMsoftware.Themis-orientationpro lesofthesedeformedareasshowcleardif-ferences(cf.Fig.3).

Inarea1,twomisorientationpro lesweredetermined,butnonerevealedthepresenceofamisorientationangleof60°,whichistypicalforatwinrelationship,andso

it

Fig.1.SEMimagesof(a)thehot-rolledmicrostructure,(b)aftercoldrolling(e=0.22).Thearrowsindicateprobablemicrotwins,(c)aftercoldrolling(e=0.51):onsetofmacroscopicshearbanding.(Imagesobtainedwithbackscatterelectroncontrast.)

高Mn合金元素TWIP钢微观组织和织构演变

L.Brackeetal./ActaMaterialia57(2009)1512–15241515

Fig.2.EBSDorientationcontrastscansoffourselecteddeformedareas.(1–3,e=0.51;4,e=0.22);colouredaccordingtoanNDinversepole gurekey.Thearrowsindicatethedirectionsinwhichthemisorientationpro lesweretaken(cf.Fig.3).

couldbeconcludedthatdeformationonlyoccurredbyslip.AR3twininfcccanbecrystallographicallydescribedbytherotationoftheparentcrystalaroundah111iaxisoveranangleof60°.Amisorientationpro lesimilartoareas1aand1bwasfoundforarea2.Itcanbeseenthatthemisori-entationpro lesshowseveralpeaksafewmicronswide,whichsuggeststhattheslipislocalized.

Themisorientationpro leinarea3clearlyshowsanumberoftwinboundaries.Thewidthofthepoint-to-ori-ginmisorientationpeaksofthetwinnedregionsisoftheorderof1lmandcorrespondstobundlesofmicrotwins.Individualmicrotwins,asobservedwithTEM(seebelow),cannotbedistinguishedwiththeEBSDequipment,butitwillbeshownthattheseverysmallmicrotwinsformthickerbandsoflocalizeddeformation,whichcanberesolved.Anotherexampleofdensebundlesofmicrotwinsispre-sentedinarea4.Itmustbenoticedthatthecold-rollingstrainforarea4wase=0.22,whileforareas1–3itwase=0.51.

Thedeformationmechanismsthatareoperativeduringcoldrollinghavebeenidenti edindetailbymeansofTEM.Fig.4showsthepresenceofmicrotwinsatarollingstraine=0.11.Dark eld(DF)imaginghasbeenusedtoclearlyidentifythetwovariantsobservedbyselectinganappropriatedi ractionspotforeachvariant.Fig.4fshowsmicrotwinsinacold-rolledsamplewithe=0.22.Thevery nemicrotwinsarestackedtogetherinarelativelythickbundleoflocalizedslip.NotethatsuchabundlecanberesolvedwiththeEBSDtechniqueused.

Itiswellknownthatmicrotwinscanbeconsideredaspeci ccon gurationofSF,e.g.,Refs.[32,33].Thetwin-ningdeformationisclearlylocalizedinbandsofnarrowlyspacedtwinlamellae.Microtwinsareknowntobestrong,impenetrablebarriersforfurtherslip[34],andthesmallspacingobservedbetweenthetwinlamellaedistinctlyreducesthemeanfreepathfordislocationslip,sothetwinsareresponsibleforthehighstrain-hardeningratesduringcoldrolling.Fig.5ashowsthatthematerialdeformsbyslip

aswell,asareaswithoutthepresenceofmicrotwinsareobserved.

Thisisexpectedingrainsthatarenotappropriatelyori-entedtoactivatetwinningsystemsandisinagreementwiththeEBSDanalysisgivenabove,whereseveraldeformedareaswerefoundfreefromtwins.

Microscopicshearbandshavebeenobservedinsamplescoldrolledtoastrainofe=0.11.ThisisillustratedintheTEMmicrographofFig.5b.Theoriginoftheseshearbandsisunclear.TheycanbeconsideredaconsequenceofoverlappingSFandmicrotwinning[9]orasaconse-quenceofplanarslipduetothelow-SFEofthematerial[33].Itispossiblethatthesemicroshearbandshaveanin uenceonthetextureevolution,aswillbediscussedlater.Althoughthedeformationenergyhasapparentlybeenaccommodatedindi erentwaysthroughoutthemicro-structure,thedeformedmicrostructureseemsenergeticallyrelativelyhomogeneous.ThisisillustratedintheKAMmapinFig.6.Thismapshowsthat,throughoutthewholedeformedmicrostructure,therearenoareaswithalargedi erenceinorientationbetweenneighbouringpoints,whichisindicativeofanenergeticallyhomogeneousmicrostructure.

3.2.Textureduringcoldrolling

TheevolutionofthetextureisshownbytheODFsec-tionsinFig.7.Thecold-rollingstrainincreasesfrome=0inFig.7atoe=0.51inFig.7f.Thehot-rolledtexture(Fig.7a)isveryweak,withamaximumintensityof2.0timesrandomforthecubecomponent{001}h100i.Thedeformationtexturesarealsorelativelyweak,withamax-imumintensityof5.0timesrandomforthesamplewitharollingstrainofe=0.51.

TheevolutionofthemaintexturecomponentswithincreasingrollingstrainisshowninFig.8.Withincreasingrollingstrain,thecubecomponentdecreasesanddisap-pearsforstrains>0.22.Theintensityofthecopper

compo-

高Mn合金元素TWIP钢微观组织和织构演变

1516L.Brackeetal./ActaMaterialia57(2009)1512–1524

Fig.3..Misorientationpro lescorrespondingtotheareasindicatedinFig.2.